Paypal, Inc.

Sedai Inc.

Ebay, Incorporated

Suresh Mathew

Benjamin Thomas

Nikhil Gopinath Kurup

Hari Chandrasekhar

Vaibhav Desai

Shankar Jothi

S Meenakshi

Aby Jacob

Krishnakanth Batta

Debashis Saha

Kumar Rethinakaleeswaran

Mahesh K Somani

David William Dodd

Ramakrishnan Sumesh Vadassery

Arvind Sugumar

Rajat Usha Krishna

Adharsh Sabukumar

Mathew Koshy Karunattu

Sayanth S

Akash Vijayan

Kumar Rethi

Rami El-Charif

Jigar Chandrakant Desai

Muthukumar Rethinakaleeswaran

Narayanan Singaram

There are provided systems and methods for a distributed autonomous patching system. During patching of a cloud computing architecture, an autonomous patching system may operator autonomously with no to minimal operator input to patch the host machines and corresponding computes of the cloud computing architecture's available applications. This may work by receiving a patch and determining corresponding patching factors for an availability zone of computes in the cloud. The system may then determine a patching topology map having an order of patching nodes for the selected computes of the application in the availability zone. Thereafter, while monitoring those computes, the system may select certain computes into an active processing funnel, where the patch is applied to those computes. Fingerprinting of before and after-patch states may be used to ensure proper patching.

Distributed autonomous patching system

Implementations described herein relate to methods, systems, and computer-readable media to provide an alert based on a release of a software application implemented in a distributed computing system. In some implementations, the method includes receiving, at a processor, an indication of the release of the software application, obtaining a first set of metric values for each metric of a list of metrics for a first time period preceding a time of release of the release, obtaining a second set of metric values for each metric of the list of metrics for a second time period following the time of release, comparing the first set of metric values to the second set of metric values to determine a deviation score, generating an alert based on the deviation score, and transmitting the alert via one of a user interface and a communication channel.

Autonomous release management in distributed computing systems

Implementations described herein relate to methods, systems, and computer-readable media to manage a computing resource allocation for a software application. In some implementations, a method may include executing a first test function using the distributed computing system at a first plurality of allocation setpoints for the computing resource, based on the execution, obtaining one or more performance metrics for the first test function for each setpoint of the first plurality of allocation setpoints, training a machine learning model based on the obtained one or more performance metrics; and utilizing the trained machine learning model to manage the computing resource for a second function.

Autonomous application management for distributed computing systems

Implementations described herein relate to methods, systems, and computer-readable media to monitor a distributed computing system. In some implementations, a method may include obtaining a first plurality of monitoring metrics of a respective application of one or more applications executing over the distributed computing system, obtaining time-series data of the first plurality of monitoring metrics, programmatically analyzing the time-series data of the first plurality of monitoring metrics to determine a second plurality of monitoring metrics, wherein the second plurality of monitoring metrics is determined to be a predictive set of the first plurality of monitoring metrics, and monitoring the distributed computing system by monitoring the second plurality of monitoring metrics.

Autonomous management of computing systems

Implementations described herein relate to run-time management of a serverless function in a serverless computing environment. In some implementations, a method includes receiving, at a processor, based on historical run-time invocation data for the serverless function in the serverless computing environment, a first number of expected invocations of the serverless function for a first time period, determining, by the processor, based on the first number of expected invocations of the serverless function for the first time period, a second number of warm-up invocation calls to be made for the first time period, and periodically invoking the second number of instances of an extended version of the serverless function during the first time period, wherein the extended version of the serverless function is configured to load and initialize the serverless function and terminate without executing the serverless function.

Autonomous concurrency for serverless functions

Implementations described herein relate to methods, systems, and computer-readable media to identify anomalous applications. In some implementations, the methods may include obtaining application metric data that includes application level metrics and instance level metrics for each instance of a plurality of instances associated with a respective application of a plurality of applications operating over a distributed computing system, generating a first anomaly detection score based on the instance level metrics; generating a second anomaly detection score based on one or more input metrics associated with the respective application, generating a third anomaly detection score based on seasonal metric data associated with the respective application and identifying at least one application of the plurality of applications as an anomalous application based on the first anomaly detection score, the second anomaly detection score, and the third anomaly detection score.

Cloud management system with autonomous aberrant behavior detection

Aspects of the present disclosure involve a system and method for autoscaling application pools. The current disclosure presents a system that can be used to understand the throughput of one instance in an application. The throughput at the one instance in the application and/or traffic trend is used to right-size or auto-scale the application pool for the current traffic trend such that the application pool can be adjusted for varying traffic trends.

Method and system for autoscaling applications

Implementations described herein relate to methods, systems, and computer-readable media to manage a computing resource allocation for a software application. In some implementations, a method may include receiving first metric data associated with the software application, determining, based on the first metric data, that an allocation of a computing resource for the software application is to be reduced from a first level of allocation, increasing allocation of the computing resource to a second level of allocation greater than the first level of allocation; executing the software application on the distributed computing system at the second level of allocation, obtaining second metric data based on execution of the software application at the second level of allocation, and reducing allocation to a third level that is lower than the first level based on a determination that the second metric data does not indicate a performance improvement for the software application.

A system and method for adaptively identifying and correcting issues in a computing system, such as a distributed node computing system, are described. The method includes receiving node data from a group of nodes, the node data describing one or more operational characteristics of a node. The operational characteristics may include CPU load, memory load, latency, or other operational data that describes node performance. Reachability data for the group of nodes is generated by trying to contact each node. Code version data is generated for each node that identifies which version of code applications in the node are running. The nodes are grouped into clusters using density-based clustering to identify outliers. A correlation is determined between the reachability, code version, and outlier data to identify problems and issue corrective actions.

Computational node adaptive correction system

With an idempotent POST request, the URL (and headers) cannot be used as an HTTP cache key. To cache idempotent POST requests, the POST body is digested and appended the URL with the digest and used as the cache key. Subsequent requests with the same payload will end up hitting the cache rather than the origin server. A forward cache proxy at the client end and reverse cache proxy at the server end are deployed. The client sends the request to the forward proxy that looks up the cache. If there is a cache miss, the forward cache proxy digests the body and sends only the digest to the reverse proxy. The reverse cache proxy looks up request cache to find if there is a match for the request and send that request to the server.

Growing community of inventors

Sunnyvale, CA, United States of America

Suresh Mathew